The first iPhone came out in 2007 and in the years since that release it seems that we have all become internet-obsessed. How much time are people actually spending connected to the internet? How do internet access and screen time vary around the world? My inspiration for this research is my reading of Ready Player One by Ernest Cline which imagines a future world where humans spend most of their time interacting virtually.

I examined data from 43 countries that assessed how much time citizens are spending connected to the internet. The data draws a distinction between mobile internet and desktop internet connection time and also includes information about the average internet speed for these two categories. Here’s a peak at how the dataset is structured:

#read in data file
screenTime<-read.csv("screenTime.csv")
#removes World Data, leaving single countries only
screenTime <- screenTime[-43,]
#renaming columns
screenTime<-screenTime %>% rename(
  mobileTime = Internet.via.Mobiles,
  mobileSpeed = Mobile.Internet.Speed..Mbps.,
  desktopTime = Internet.via.Computer,
  desktopSpeed = Fixed.Internet.Speed..Mbps.,
  totalScreenTime = Total.Time.Spent.on.Devices)
head(select(screenTime,Country,totalScreenTime,mobileTime,desktopTime,mobileSpeed,desktopSpeed))
##     Country totalScreenTime mobileTime desktopTime mobileSpeed desktopSpeed
## 1 Argentina           09:38      05:04       04:34       21.09        38.61
## 2 Australia           06:13      02:49       03:24       81.14        51.98
## 3  Austria            05:42      02:29       03:13       50.31        51.01
## 4   Belgium           05:44      02:19       03:25       47.98        77.58
## 5    Brazil           10:19      05:25       04:54       22.60        83.25
## 6   Canada            06:45      02:51       03:54       72.87        97.51

It turns out that the average American spends seven hours and four minutes online each day, nearly 44% of their waking hours! They spend three hours and 34 minutes using desktop internet and three hours and 30 minutes using mobile internet. Two hours and 14 minutes of this time on average is spent on social media. 92% of Americans use the internet daily, but how do we compare to residents of other countries around the world? Consider India where just 47% of citizens are online daily! Japan, a well-developed, technological society averages only four hours and 25 minutes of daily internet use! Is the United States in the top 10 around the world for daily internet-connected screen time?

#barplot
#top screen time countries
#Japan: only 4:25 average screen time
#Sort screen times in descending order:
screenTimeDesc <- screenTime[order(screenTime$totalScreenTime,decreasing=TRUE),]
#selects top 10 countries
screenTimeTop10 <- head(screenTimeDesc,10)
#slices out number of hours
top10Hrs<-cbind(substring(screenTimeTop10$totalScreenTime,1,2))
#slices out number of minutes
top10Min<-cbind(substring(screenTimeTop10$totalScreenTime,4,5))
#converts to numeric data type
top10HrsNum<-as.numeric(top10Hrs)
top10MinNum<-as.numeric(top10Min)
#calculates total number of minutes
top10HrsNumMin<-top10HrsNum*60
top10HrsInMins<-top10HrsNumMin + top10MinNum
#new column with hours rounded to one decimal place
screenTimeTop10$top10Hrs<-round(top10HrsInMins/60, digits = 1)
#editing longer country names
screenTimeTop10$Country[1]="Sth Africa"
screenTimeTop10$Country[10]="UAE"
#barplot in descending order with color and labels
ggplot(screenTimeTop10, 
       aes(x = reorder(Country, -top10Hrs), 
           y = top10Hrs)) + 
  geom_bar(stat="identity",fill = "lightblue", color="darkblue")+
  labs(x = "Country", 
       y = "Daily Screen Time (hrs)", 
       title = "Largest Daily Screen Time by Country")

The US doesn’t make the top ten, instead ranking twenty-first in the world. It must be noted that not all countries have equal access to the internet and quality of connection varies widely around the world. Colombia, for example, has an average mobile internet speed of just 13.34 Mbps. Compare this with the US which has an average desktop internet speed of 134.10 Mbps and mobile internet speed of 53.31 Mbps. How different is internet speed around the world? Where does the US fall in the world distribution?

#boxplot
#internet speeds differ around world (Colombia only 13.34 Mbps mobile internet speed)
#How much difference is there in speeds?
#US Desktop Speed: 134.10 Mbps
#US Mobile Speed: 53.31 Mbps
#selects relevant columns
netSpeed<-select(screenTime, mobileSpeed, desktopSpeed)
#wrangles mobile and desktop speeds into one column
screenAndMobile<-c(netSpeed[,1],netSpeed[,2])
#creates a new column with Mobile and Desktop labels
labels = c(rep("Mobile",times=43),
           rep("Desktop",times=43))
#combines the labels and numbers
bpTable <- cbind(labels,screenAndMobile)
#makes it into a data frame
bpTableDF <- as.data.frame(bpTable)
#converts speed numbers to numeric form
bpTableDF$screenAndMobile <- as.numeric(bpTableDF$screenAndMobile)
#boxplot with mobile and desktop internet speeds with appropriate labels
ggplot(bpTableDF, aes(x=labels, y=screenAndMobile))+
  labs(title="Internet Speed Distribution by Connection Type",
       x="Internet Connection Type",
       y="Mobile Internet Speed (Mbps)")+
  geom_boxplot()

It is interesting to note that mobile internet speeds on the whole are slower than desktop speeds around the world. The US is a bit off the world pace for mobile internet speed finding itself in the second quartile. The US is near the top of the group in desktop internet speed. Internet speeds vary around the world and this leads one to wonder if a slower connection could be a reason for increased screen time. Consider these two scatterplots examining mobile and desktop internet usage by internet connection speed.

#2 scatterplots
#does internet speed affect screen time (source says no)?
#plot mobile and desktop speed to screen time
#selects number of hours and minutes for mobile
mobileHrs<-cbind(substring(screenTime$mobileTime,1,2))
mobileMin<-cbind(substring(screenTime$mobileTime,4,5))
#converts to numeric type
mobileHrsNum<-as.numeric(mobileHrs)
mobileMinNum<-as.numeric(mobileMin)
#calculates total number of minutes
mobileHrsNumMin<-mobileHrsNum*60
mobileHrsInMin<-mobileHrsNumMin + mobileMinNum
#adds a column with this information
screenTime$mobileMins<-mobileHrsInMin
#scatterplot 1 with mobile screen time by speed with a regression line and labels
mobileTimeSpeed<-ggplot(data = screenTime,
       aes(x = mobileSpeed, y = mobileMins))+
  geom_smooth(method="lm",se=FALSE,linewidth=1,formula = y ~ x)+
  geom_point()+
  labs(title = "Mobile Screen Time by Connection Speed",
       x = "Mobile Connection Speed (Mbps)",
       y = "Daily Mobile Screen Time (mins)"
)
ggplotly(mobileTimeSpeed)
#selects number of hours and minutes for desktop internet
desktopHrs<-cbind(substring(screenTime$desktopTime,1,2))
desktopMin<-cbind(substring(screenTime$desktopTime,4,5))
#converts to numeric type
desktopHrsNum<-as.numeric(desktopHrs)
desktopMinNum<-as.numeric(desktopMin)
#calculates total number of minutes
desktopHrsNumMin<-desktopHrsNum*60
desktopHrsInMins<-desktopHrsNumMin + desktopMinNum
#adds a column with this information
screenTime$desktopMin<-desktopHrsInMins
#scatterplot 2 with desktop screen time by speed with regression line and labels
fixedTimeSpeed<-ggplot(data = screenTime,
        aes(x = desktopSpeed, y = desktopHrsInMins))+
        geom_smooth(method="lm",se=FALSE,linewidth=1,formula = y ~ x)+
        geom_point()+
  labs(title = "Desktop Screen Time by Connection Speed",
       x = "Desktop Connection Speed (Mbps)",
       y = "Daily Mobile Screen Time (min)"
)
ggplotly(fixedTimeSpeed)

Though my source disagrees, my visualizations show a weak, inverse correlation between connection speed and average screen time.

The world wide web has rapidly changed humanity’s daily routines. We now spend a significant portion of our days connected to the internet. How this behavioral shift will affect our species remains to be seen but it deserves to be examined thoroughly.

The inspiration for this project came from an article written by Rebecca Moody, Head of Data Research at Comparitech. The dataset was downloaded from: https://www.comparitech.com/tv-streaming/screen-time-statistics/ which sources the information to https://datareportal.com/